mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-07 12:21:27 +01:00
8d45f555d7
Rewrite Python built-in class `super()` calls. Only non-semantic changes should be applied. - #94587 - #94588 - #94592 Also, methods with only a `super()` call are removed: ```diff class MyModule(nn.Module): - def __init__(self): - super().__init__() - def forward(self, ...): ... ``` Some cases that change the semantics should be kept unchanged. E.g.:f152a79be9/caffe2/python/net_printer.py (L184-L190)f152a79be9/test/test_jit_fuser_te.py (L2628-L2635)Pull Request resolved: https://github.com/pytorch/pytorch/pull/94587 Approved by: https://github.com/ezyang
63 lines
1.9 KiB
Python
63 lines
1.9 KiB
Python
## @package random_neg_rank_loss
|
|
# Module caffe2.python.layers.random_neg_rank_loss
|
|
|
|
|
|
|
|
|
|
|
|
from caffe2.python import schema, core
|
|
from caffe2.python.layers.layers import (
|
|
ModelLayer,
|
|
)
|
|
from caffe2.python.layers.tags import (
|
|
Tags
|
|
)
|
|
import numpy as np
|
|
|
|
|
|
class MarginRankLoss(ModelLayer):
|
|
|
|
def __init__(self, model, input_record, name='margin_rank_loss',
|
|
margin=0.1, average_loss=False, **kwargs):
|
|
super().__init__(model, name, input_record, **kwargs)
|
|
assert margin >= 0, ('For hinge loss, margin should be no less than 0')
|
|
self._margin = margin
|
|
self._average_loss = average_loss
|
|
assert schema.is_schema_subset(
|
|
schema.Struct(
|
|
('pos_prediction', schema.Scalar()),
|
|
('neg_prediction', schema.List(np.float32)),
|
|
),
|
|
input_record
|
|
)
|
|
self.tags.update([Tags.EXCLUDE_FROM_PREDICTION])
|
|
self.output_schema = schema.Scalar(
|
|
np.float32,
|
|
self.get_next_blob_reference('output'))
|
|
|
|
def add_ops(self, net):
|
|
neg_score = self.input_record.neg_prediction['values']()
|
|
|
|
pos_score = net.LengthsTile(
|
|
[
|
|
self.input_record.pos_prediction(),
|
|
self.input_record.neg_prediction['lengths']()
|
|
],
|
|
net.NextScopedBlob('pos_score_repeated')
|
|
)
|
|
const_1 = net.ConstantFill(
|
|
neg_score,
|
|
net.NextScopedBlob('const_1'),
|
|
value=1,
|
|
dtype=core.DataType.INT32
|
|
)
|
|
rank_loss = net.MarginRankingCriterion(
|
|
[pos_score, neg_score, const_1],
|
|
net.NextScopedBlob('rank_loss'),
|
|
margin=self._margin,
|
|
)
|
|
if self._average_loss:
|
|
net.AveragedLoss(rank_loss, self.output_schema.field_blobs())
|
|
else:
|
|
net.ReduceFrontSum(rank_loss, self.output_schema.field_blobs())
|